Austin PC Optimal caliper widths for propensity
20. Austin PC. Optimal caliper widths for propensity-score matching when estimating differences in means and differences in proportions in observational studies. Pharma Stat. 2011;10:150–161.
22. Tosoian JJ, Mamawala M, Epstein JI, et al. Intermediate and longer-term outcomes from a prospective active-surveillance program for favorable-risk prostate cancer. J Clin Oncol. 2015;33:3379–3385.
23. Leapman MS, Cowan JE, Nguyen HG, et al. Active surveillance in younger men with prostate cancer. J Clin Oncol. 2017;35:1898–1904. 24. Ouzzane A, Renard-Penna R, Marliere F, et al. Magnetic resonance imaging targeted biopsy improves selection of patients considered for active surveillance for clinically low risk prostate cancer based on systematic biopsies. J Urol. 2015;194:350–356. 25. Muthigi A, Sidana A, George AK, et al. Current beliefs and practice patterns among urologists regarding prostate magnetic resonance imaging and magnetic resonance-targeted biopsy. Urol Oncol. 2017;35. 32.e31-32.e37.
28. Schmid M, Meyer CP, Reznor G, et al. Racial differences in the sur-gical care of medicare beneficiaries with localized prostate cancer. JAMA Oncol. 2016;2:85–93. 29. Davis BA, Aminawung JA, Abu-Khalaf MM, et al. Racial and eth-nic disparities in oncotype DX test receipt in a statewide population-based study. J Natl Compr Cancer Network. 2017;15:346–354.
30. Gordon LG, James R, Tuffaha HW, Lowe A, Yaxley J. Cost-effec-tiveness analysis of multiparametric MRI with increased active sur-veillance for low-risk prostate cancer in Australia. JMRI. 2017;45:1304–1315.
Active surveillance and prostate magnetic resonance imaging (MRI) have revolutionized the management of prostate cancer in recent years. This study evaluates the association between the use of prostate MRI and observation for 8144 men with newly diagnosed low-risk prostate cancer using data from the Surveillance, Epidemiology, and End Results-Medicare data-base. From 2010-2013, the use of MRI increased 3-fold and the use of observation increased from 30.8%-48.1% in these patients. After propensity-score matching to control for poten-tial confounders captured in the data, the authors demonstrate that GSK1838705A men who received a prostate MRI surrounding the diag-nosis of prostate cancer were significantly more likely to undergo observation. The authors conclude that prostate MRI may increase confidence in the assignment of low-risk classifi-cation and thereby facilitate the use of active surveillance for prostate cancer.
This study offers real-world data that may support the beliefs of many urologists: a prostate MRI without suspicious lesions can make a strong case for active surveillance in an otherwise appropriate candidate. However, there are alterna-tive explanations for the association between MRI and obser-vation noted in this study and the authors are appropriately cautious about making an argument for a causal relationship. Prostate MRIs captured in this analysis may have been used in men who were already being managed with active surveil-lance. In that context, the use of observation “caused” the MRI more than the converse. Alternatively, there may be unmeasured confounders that are associated with both the use of observation and the use of prostate MRI. One such possibil-ity is a characteristic of the treating physician. Physicians who are more likely to order a prostate MRI for their patients may also be more likely to recommend observation for men with low-risk disease. Particularly in the study period (2010-2013), physicians who were earlier adopters of prostate MRI may also have been more likely to recommend active surveillance to their low-risk prostate cancer patients. Propensity-score matching, which the authors used to generate matched cohorts of patients with and without prostate MRI, can only account for covariates that are captured in the data and can-not control for physician- or hospital-factors that are not available in administrative claims.1
Understanding the use of active surveillance on a national level is critically important for the field of urology. Despite long-term data from several centers supporting the safety of active surveillance, its use still varies considerably from physi-cian to physician2 and the optimal protocol remains unknown.3,4 Innovations such as prostate MRI and biomarker tests offer us the potential to further refine our patient selec-tion, but we do not know exactly how these tools are being used. Analyses such as this one, using nationally representa-tive data, may help us better understand how these pieces fit
together and how we might continue to improve the manage-ment of men with prostate cancer.
ParthX1D K. Modi, MD,X2D MS, Department of Urology, Dow Division of Health Services Research, Michigan Medicine, Ann Arbor, MI